Can FSS Code CO2 Systems Extinguish a Li-Ion Fire at Sea?

The Fire Safety Systems Code mandates a minimum 45% CO2 concentration by volume for enclosed ro-ro vehicle spaces — sufficient to starve hydrocarbon combustion of atmospheric oxygen. A lithium-ion cell in thermal runaway generates its own oxygen through internal electrochemical decomposition. No atmospheric CO2 concentration achievable in a cargo hold can stop a cell past Stage 2 of thermal runaway. The question is not whether CO2 works; it is what CO2 can and cannot defend against.

How CO2 suppresses a conventional vehicle fire

Hydrocarbon combustion — fuel-air oxidation in a conventional vehicle fire — requires atmospheric oxygen above roughly 15% by volume. CO2 flooding displaces air until that threshold is crossed, smothering the reaction. The FSS Code Chapter 5 sets the design quantity for enclosed ro-ro and vehicle spaces at a minimum gas volume equal to 45% of the gross volume of the largest sealed cargo space, with at least two-thirds of that quantity discharging within 10 minutes. The 45% figure carries a headroom margin above the 30–35% typically required to extinguish Class B hydrocarbon fires, accounting for leakage and incomplete sealing in enclosed decks.

Critically, this mechanism is entirely dependent on removing the atmospheric oxygen supply. It works because the fuel and the oxygen are two separate variables — withdraw the oxygen and the reaction stops.

Why a Li-ion cell in thermal runaway ignores atmospheric oxygen

A lithium-ion cell at Stage 3 or beyond is not a fuel-air system. The cathode material — LiCoO₂, NMC, or similar — undergoes exothermic decomposition at temperatures typically above 150–200°C, releasing oxygen from within the crystal structure itself. The electrolyte, an organic solvent such as ethylene carbonate or dimethyl carbonate combined with a lithium salt, provides the fuel in the same enclosed volume. The oxidiser and the fuel are both inside the cell, separated by a membrane that the heat eventually defeats. By Stage 3, that separation has failed.

The consequence is direct: the atmospheric oxygen concentration in the hold — whether 21% (air) or 0% (fully flooded with CO2) — is irrelevant to the internal reaction in the failing cell. CO2 cannot reach inside the cell and cannot displace the internally-sourced oxygen driving the cathode decomposition. The reaction continues regardless.

What CO2 actually defends against — and where its limit lies

The correct framing is not "CO2 fails on EV fires." CO2 performs its designed function: it suppresses secondary combustion — the burning of emitted gases (H2, CO, C₂H₄, and others) and the fire spreading from the origin vehicle to adjacent vehicles through atmospheric combustion. On a PCTC with thousands of vehicles in undivided horizontal rows, preventing that vehicle-to-vehicle cascade is not a minor function. It is the difference between losing one vehicle and losing the ship.

The limit is this: CO2 must be activated before the fire load grows large enough that cascade propagation is already under way. Once 60–70 vehicles are simultaneously burning — as in the MV Delphine (Zeebrugge, April 2025) — CO2 contains rather than prevents. Each additional burning vehicle adds its own internal oxygen source to the deck.

The re-ignition problem

A second limitation is temporal. Conventional vehicle fires cool when starved of oxygen and do not spontaneously reignite once CO2 concentration drops. EV batteries can reignite hours or even days after a blaze appears extinguished, as IUMI notes in its September 2025 best practice paper. The internal electrochemical state of a partially-failed cell — residual charge, mechanical deformation, continued electrolyte decomposition — can restart thermal runaway long after the visible fire is out and CO2 has dissipated. The Zeebrugge fire department held nitrogen tanks on standby for reflash prevention after the Delphine fire was declared contained. That is not standard protocol for a conventional vehicle fire.

Why IUMI recommends doubling CO2 capacity on PCTCs

The FSS Code 45% design concentration was developed for ICE-era cargo. IUMI's September 2025 updated best practice paper states that CO2 firefighting capacity on PCTCs should be at least doubled. The engineering rationale is not that doubling changes the suppression mechanism — it does not make CO2 effective against a self-oxidising cell — but that it addresses two EV-specific failure modes of the current sizing.

• Duration: EV fires burn longer and at higher peak temperatures (>1,000°C vs ~600°C for ICE), requiring a sustained CO2 supply to maintain the concentration needed to suppress secondary combustion while the cell exhausts itself.
• Dilution: thermal runaway off-gases H2, CO, and other products at volume into the enclosed hold, actively diluting the CO2 concentration mid-discharge. A system sized only for the 45% target may fall below that threshold before the fire is suppressed.
• Re-ignition: subsequent thermal events from the same or adjacent cells require a second CO2 discharge. Current PCTCs carry one discharge quantity; the second event has no CO2 left.

IUMI also identifies total-flood high-pressure water mist as an alternative for PCTCs, noting its effectiveness in preventing heat transfer between vehicles. Water mist directly addresses the EV fire's key limitation — it cools the cell body, slowing or stopping the internal decomposition — but requires a different structural and plumbing approach than existing CO2 system infrastructure.

Detection timing is what makes CO2 effective or ineffective

The Delphine case is instructive. CO2 was activated; the fire was brought under control as designed; 60–70 vehicles had already burned. That is the expected outcome when CO2 activates at the smoke-detection stage — Stage 3 is already complete in the origin vehicle, and cascade to neighbouring vehicles has begun. A system that catches the electrochemical precursors of thermal runaway at Stage 1 or Stage 2 changes the CO2 activation timing: CO2 fires into a deck where one cell is in distress, not one where sixty vehicles are already involved.

Earlier activation does not change what CO2 can do at the cell level. It changes the fire load the CO2 is defending against — and at a small enough fire load, containment and prevention become the same outcome.

Sources

• IMO Fire Safety Systems (FSS) Code, Chapter 5 — adopted by Resolution MSC.98(73). CO2 system quantity for ro-ro and vehicle spaces: 45% of gross volume, two-thirds within 10 minutes.
• IUMI — "Risk mitigation for the safe ocean and short-sea carriage of electric vehicles" (September 2025). CO2 doubling recommendation; >1,000°C peak temperature; re-ignition window.
• Maritime Executive — "Major EV Fire Breaks Out Aboard Ro/Ro at Zeebrugge" (April 2025). CO2 system activated, 60–70 vehicles destroyed.
• [VERIFY: Peer-reviewed citation for Li-ion cathode decomposition yielding internal O₂ at temperature (e.g. LiCoO₂ → Li₁₋ₓCoO₂ + O₂). Suggested source: Feng et al., "Thermal runaway mechanism of lithium ion battery for electric vehicles", Energy Storage Materials (2018), or equivalent. Confirm exact decomposition onset temperatures before publication.]
• [VERIFY: Confirm that the FSS Code 45% figure is in Chapter 5 (CO2 systems) rather than another chapter, against the IMO source above. The search result cited Chapter 5, regulation 2; verify section numbering.]